Context

This report evaluates the results of an analytical run according to the EMA Guideline on bioanalytical method validation (Rev.1 Corr.2, 2011). It does not substitute the analytical- or the validation report as key information on the used protocol is not incorporated (e.g. storage conditions, batch numbers, …).
The script was written in R 4.1.0 and formatted with the aid of the rmarkdown package. Documentation and source code of the script can be found at the Github repository of the original author. The report generator can be cited as: De Sutter PJ, Bio-analytical report generator, (2022), GitHub repository, https://github.com/PJDeSutter/Method-Validation

1.1. Procedure

1.2. Results

2.1. Definition

“The calibration curve is the relationship between instrument response and known concentration of the analyte.”

2.2. Acceptance criteria

“The back calculated concentrations of the calibration standards should be within ±15% of the nominal value, except for the LLOQ for which it should be within ±20%. At least 75% of the calibration standards, with a minimum of six, must fulfill this criterion.”
“In case a calibration standard does not comply with these criteria, this calibration standard sample should be rejected, and the calibration curve without this calibration standard should be re-evaluated, including regression analysis”

2.3. Results

2.4. Conclusion

• Number of calibrators outside the applicable tolerance ranges for back-calculated concentration: 0 ( 0% )
  • At least 75% of the calibrators fall inside the tolerance ranges? –> YES
  • At least 6 calibrators fall inside the tolerance ranges –> YES
• Based on these criteria the calibration curve is accepted

3.1 Definition

“Accuracy is the closeness of the determined concentration to the nominal concentration.”

\[ Accuracy\ (bias)= \frac{Concentration_{fitted} - Concentration_{nominal}}{Concentration_{nominal}}*100\]

3.2. Acceptance criteria

Analytical run: Accuracy of the mean fitted concentration should fall within ±15% bias of the nominal values. At least 67% of the QC samples and at least 50% at each concentration level should comply with this criterion.
Method validation: within-run accuracy: minimum of 5 samples at a minimum of 4 concentration levels covering the calibration curve range (see section 5.3). The mean concentration should be within 15% of the nominal values for the QC samples, except for the LLOQ which should be within 20% of the nominal value
Method validation: between-run accuracy: LLOQ, low, medium, high QC samples from at least three runs analysed on at least two different days should be evaluated. The mean concentration should be within 15% of the nominal values for the QC samples, except for the LLOQ which should be within 20% of the nominal value.

3.3. Results

3.4. Conclusion (analytical run)

• Amount of QC samples outside the applicable tolerance ranges for accuracy: 0 ( 0% )
  • At least 67% of the QC samples complies with tolerance ranges for accuracy? –> YES
  • At least 50% of the QC samples at each QC level complies with the tolerance ranges for accuracy? –> YES
• Amount of QC levels outside the applicable tolerance ranges for accuracy: 0 ( 0% )
  • All QC levels fall inside the applicable tolerance ranges for accuracy ? –> YES
• Based on these 3 accuracy criteria the accuracy of the QC samples is accepted

4.1. Definition

“Precision is the closeness of repeated individual meassures of the analyte.”

\[ Precision\ (\% CV)= \frac{Standard\ deviation_{fitted\ concentration }}{Mean_{fitted\ concentration}}*100\]

4.2. Acceptance criteria

Analytical run: The coefficient of variation (CV) on the replicates should be less than 15% CV for each QC level.
Method validation: within-run precision: minimum of 5 samples at a minimum of 4 concentration levels covering the calibration curve range (see section 5.3) in a single run. The within-run CV value should not exceed 15% for the QC samples, except for the LLOQ which should not exceed 20%
Method validation: between-run accuracy: LLOQ, low, medium, high QC samples from at least three runs analysed on at least two different days should be evaluated. The between-run CV value should not exceed 15% for the QC samples, except for the LLOQ should not exceed 20% of the nominal value.

4.3. Results

4.4. Conclusion (analytical run)

• Amount of QC levels exceeding applicable tolerance ranges for precision: 0 ( 0% )
  • All QC levels confirm with applicable criteria for precision ? –> YES
• Based on these precision criteria the analytical run is accepted

5.1 LLOQ (lowest standard)

5.2. ULOQ (highest standard)

5.3. QC samples

6.1. Analytes: all study samples

6.2. Analytes: average of sample replicates

6.3. Analtyical run criteria

1. At least one blank sample (processed matrix sample without analyte and without IS)?
   • YES (7 blank samples)
2. At least one zero sample (processed matrix sample with IS)?
   • NO (0 zero samples)
3. At least 6 calibration levels?
   • YES (8 calibrator levels)
4. At least 3 QC levels?
   • YES (4 QC levels)
5. QC samples at least in duplicate or 5% of the analytical samples (whichever is greater)?
   • YES (12 QC samples)
     • 5% of the analytical samples = 0.25
     • QC levels in duplicate = 8

6.4. Distribution of analytical samples

1. At least 2 QC sample levels within the range of concentrations measured in study samples ?
   • Evaluate visually if study-samples are analysed

7.1. Internal standard

7.2. Peak symmetry

7.3. Retention time

7.4. S/N ratio

7.5. Quantifier/Qualifier ratio

7.6. Calibration model

## 
## Call:
## lm(formula = Std_conc_mean ~ Response_mean, data = calsum, weights = 1/(calsum$Std_conc_mean^2))
## 
## Weighted Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.085522 -0.042979 -0.006508  0.040080  0.140982 
## 
## Coefficients:
##                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   0.0001513  0.0022782   0.066    0.949    
## Response_mean 0.5240830  0.0196821  26.627 1.85e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.08478 on 6 degrees of freedom
## Multiple R-squared:  0.9916, Adjusted R-squared:  0.9902 
## F-statistic:   709 on 1 and 6 DF,  p-value: 1.852e-07

##  lag Autocorrelation D-W Statistic p-value
##    1       0.4696953     0.9212717   0.014
##  Alternative hypothesis: rho != 0

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

7.7. Other weighing functions

7.7.2. QC samples with alternative weighing functions

QC samples: Unweighed regression model

Level	Replicate	Nominal concentration (µg/mL)	Fitted concentration (µg/mL)	Accuracy (%)	Passes accuracy criterium?
LLOQ	1	0.03	0.04	64.01	FALSE
LLOQ	2	0.03	0.05	94.23	FALSE
LLOQ	3	0.03	0.05	83.50	FALSE
LOW	1	0.04	0.06	49.37	FALSE
LOW	2	0.04	0.06	55.08	FALSE
LOW	3	0.04	0.06	56.47	FALSE
MID	1	0.41	0.44	6.40	TRUE
MID	2	0.41	0.43	4.13	TRUE
MID	3	0.41	0.46	12.04	TRUE
HIGH	1	2.32	2.06	-11.21	TRUE
HIGH	2	2.32	2.44	5.17	TRUE
HIGH	3	2.32	2.51	7.95	TRUE
a concentration units = µg/mL

Precision QC samples, unweighed model

Level	Nominal concentration	Mean fitted concentration	SD fitted concentration	Precision (%)	Passes precision criterium?
LLOQ	0.03	0.05	0.00	8.48	TRUE
LOW	0.04	0.06	0.00	2.45	TRUE
MID	0.41	0.44	0.02	3.79	TRUE
HIGH	2.32	2.34	0.24	10.29	TRUE
a concentration units = µg/mL

QC samples: 1/x weighed regression model

Level	Replicate	Nominal concentration (µg/mL)	Fitted concentration (µg/mL)	Accuracy (%)	Passes accuracy criterium?
LLOQ	1	0.03	0.03	0.89	TRUE
LLOQ	2	0.03	0.03	31.68	FALSE
LLOQ	3	0.03	0.03	20.75	FALSE
LOW	1	0.04	0.05	10.81	TRUE
LOW	2	0.04	0.05	16.63	FALSE
LOW	3	0.04	0.05	18.05	FALSE
MID	1	0.41	0.43	4.25	TRUE
MID	2	0.41	0.42	1.93	TRUE
MID	3	0.41	0.45	9.99	TRUE
HIGH	1	2.32	2.08	-10.29	TRUE
HIGH	2	2.32	2.47	6.40	TRUE
HIGH	3	2.32	2.54	9.23	TRUE
a concentration units = µg/mL

Precision QC samples, 1/x weighed model

Level	Nominal concentration	Mean fitted concentration	SD fitted concentration	Precision (%)	Passes precision criterium?
LLOQ	0.03	0.03	0.00	13.25	TRUE
LOW	0.04	0.05	0.00	3.33	TRUE
MID	0.41	0.43	0.02	3.94	TRUE
HIGH	2.32	2.36	0.24	10.36	TRUE
a concentration units = µg/mL

8.1. Raw Data

Version A.3.

Creator Rmarkdown script: Pieter-Jan De Sutter (Ghent University)